Skip to Main Content
 

Global Search Box

 
 
 
 

Files

File List

Full text release has been delayed at the author's request until March 31, 2026

ETD Abstract Container

Abstract Header

SPECTRAL DECOMPOSITION AND MODEL ESTIMATION OF SOIL CARBON IN SITES WITH MULTIPLE SOIL TYPES

Baimah, Mohammed
http://orcid.org/0000-0001-6723-8214
http://rave.ohiolink.edu/etdc/view?acc_num=kent1745875080965517

Abstract Details

2025, MS, Kent State University, College of Arts and Sciences / Department of Earth Sciences.
Remote sensing and watershed modeling are a rapid and non-invasive approach to monitoring agricultural systems in the current climate crisis with a shortage of food production in many areas. The soil matrix is very complex with varying land cover and soil materials, making it difficult to identify unique spectral curves and model specific soil constituents. This study seeks to estimate soil carbon in farms with multiple soil types based on model simulations and the decomposition of visible and near-infrared spectral response signals. First, the current study applied Kent State University’s Varimax-Rotated Principal Component Analysis (KSU VPCA) through the Semi-Automated Averaging Code to unmix groundcover spectral signals and detect carbon, and nitrogen. The KSU VPCA of a reflectance obtained from the FieldSpec4 SR and the Harmonized Sentinel-2A/B imagery through stepwise multiple regression explained variance of 45-68% in soil percentage of total carbon (R = 0.68-0.83, and an R2 =0.45-0.68), 45-67% in soil percentage of total nitrogen (R = 0.67-0.82, and an R2 =0.45-0.67), and 16-40% in permanganate oxidizable carbon (R = 0.29-0.63, and an R2 = 0.16-0.40). The six-eigen vector KSU VPCA correlated strongly with spectral signals of a mixture of minerals and plant-laden pigments. The KSU VPCA of plant materials explained an 81% variance in crop vegetation biomass. To predict groundcover, Random Forest Classifiers (RFC) were developed based on the first derivatives of reflectance (RFC2), and KSU VPCA scores of reflectance (RFC3). RFC2 predicted dried crop residues with the highest precision of 100% while RFC3 predicted soil with the highest recall score of 95%. Secondly, the study applied the Agricultural Policy eXtender model to assess the impacts of cover crops and tillage on soil carbon stocks in a maize -wheat-soybean crop rotation. The reduced tillage with cover crops improved soil organic carbon (50-110%) compared to conventional tillage (5-48%). A tri-state fertility farming practice scenario resulted in 40-115% more soil organic carbon but produced unrealistic yields. Thus, the results imply that the KSU VPCA can be used as a proxy for measuring carbon stocks and the carbon stocks can be improved with the continuous application of reduced tillage and cover cropping. The results are critical for precision agricultural operations, hyperspectral ground truthing, and large-scale carbon monitoring.
Joseph Ortiz (Advisor)
Timothy Gallagher (Committee Member)
Sarah Eichler (Committee Member)
266 p.
Geology
Kent State University’s Varimax-Rotated Principal Component Analysis; Agricultural Policy eXtender model; Random Forest; Soil Carbon; Sentinel 2; Hyperspectral; Multispectral; Biomass; Spectral Decomposition; Satellite; Remote Sensing; Spectral; Multiple Regression; Derivatives

Recommended Citations

Citations

  • Baimah, M. (2025). SPECTRAL DECOMPOSITION AND MODEL ESTIMATION OF SOIL CARBON IN SITES WITH MULTIPLE SOIL TYPES [Master's thesis, Kent State University]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=kent1745875080965517

    APA Style (7th edition)

  • Baimah, Mohammed. SPECTRAL DECOMPOSITION AND MODEL ESTIMATION OF SOIL CARBON IN SITES WITH MULTIPLE SOIL TYPES . 2025. Kent State University, Master's thesis. OhioLINK Electronic Theses and Dissertations Center, http://rave.ohiolink.edu/etdc/view?acc_num=kent1745875080965517.

    MLA Style (8th edition)

  • Baimah, Mohammed. "SPECTRAL DECOMPOSITION AND MODEL ESTIMATION OF SOIL CARBON IN SITES WITH MULTIPLE SOIL TYPES ." Master's thesis, Kent State University, 2025. http://rave.ohiolink.edu/etdc/view?acc_num=kent1745875080965517

    Chicago Manual of Style (17th edition)

kent1745875080965517
© 2025, all rights reserved.
This open access ETD is published by Kent State University and OhioLINK.
Release 3.2.12